Introduction Results

advertisement
Crown Dieback Trends Across the Southeastern United States
Michael K. Crosby1, Zhaofei Fan1, Martin A. Spetich2, Theodor D. Leininger3, and Xingang Fan4
1Mississippi
State University – Forest and Wildlife Research Center, 2U.S. Forest Service - Arkansas Forestry Sciences Laboratory 3U.S. Forest Service – Center for
Bottomland Hardwoods Research, 4Western Kentucky University – Department of Geography and Geology
Introduction
Results
• Crown health indicators, specifically estimates of crown
dieback, are provided by the U.S. Forest Service’s Forest
Health Monitoring (FHM) program and provide a means
of assessing forested areas that may be at risk for
subsequent damage
• Crown dieback estimates, in 5% categories, are obtained
for all trees on Phase 3 inventory plots
• Extreme fluctuations in climatic variables (i.e., Palmer’s
Drought Severity Index (PDSI)) can cause damage to
healthy forests or exacerbate stress on trees already
impacted by other events (e.g., severe weather, insect
outbreak, etc.)
• Impacts of drought are not uniform for all species groups
which could lead to difficulty in developing models to
assess the impact of drought throughout the southeastern
United States
• Red oak species group shows the highest relative crown
dieback occurring throughout the study area, particularly
in south-central Alabama, the Ozark Highlands in
northern Arkansas, north Georgia, and eastern Tennessee
(Figure 1(a))
• Red oak CART model shows the first split occurring with
forest type and then by ecoregion with 1-year, 3-year, and
7-year cumulative drought being the most significant
drought variables (Figure 2)
• The white oak species group exhibits the highest levels of
crown dieback in eastern Texas, west/northwestern
Arkansas, and areas along the coast of Florida
Figure 1(b))
• The CART model shows the highest average relative
crown dieback for white oaks occurring in the Prairie
Parkland ecoregion, in pine, oak/hickory, and
oak/gum/cypress forest types, with a low 7-year
cumulative PDSI value (Figure 3)
• The highest values of relative crown dieback in the other
hardwoods species group occurs primarily along coastal
Alabama and Florida and eastern Texas (Figure 1(c))
• CART results indicate that the most significant split for
crown dieback occurs in the Pine and Oak/Pine forest
types (Figure 4)
(a)
(b)
Figure 1. Kernel smoothing results for (a) red oak species group, (b) white oak species group, and (c) other hardwoods species group.
Objectives
1. Investigate the spatial trend in crown dieback across the
southeastern United States for red oak, white oak, and
other hardwoods species groups.
2. Determine the relationship between crown dieback and
drought by Ecoregion and forest type using Classification
and Regression Tree (CART) analysis.
(c)
Figure 2. Classification and Regression Tree for the red oak species group.
Methodology
Conclusions
• Obtained Phase 3 FIA/FHM data for 10 states in the
southeastern U.S. (Alabama, Arkansas, Florida, Georgia,
Kentucky, North Carolina, South Carolina, Tennessee,
Texas, and Virginia) for the years 2002-2010
• Obtained PDSI data from the National Climate Data
Center (NCDC) and calculated cumulative growing season
PDSI (March-September) for a 10 year period prior to
inventory years
• Used ArcGIS to extract annual growing season PDSI and
Ecoregion to each FHM plot
• Determined relative crown dieback for each species group
by calculating an Relative Crown Dieback variable which
removes variations due to stand and site factors:
• Spatially, all species groups show elevated levels of
crown dieback occurring along coastal sections of
Alabama and Florida and areas of eastern Texas possibly
owing to prolonged drought or storm damage (with
secondary impacts from insects)
• CART results indicate that drought plays a role in higher
levels of crown dieback as well as classification by
ecoregion and/or forest type
• These preliminary results can be utilized to develop
hierarchical models to assess the current future impacts of
climate change scenarios based on downscaled climate
data
Figure 3. Classification and Regression Tree for the white oak species group.
Average Crown Dieback for Species Group x i
Relative Crown Dieback =
Total Crown Dieback for all Species Groups
• Used Kernel Density to determine spatial trends for each
species group across the selected states
• Developed CART models for each species group that split
the data into homogenous groups based on forest type,
ecoregion, and cumulative growing season PDSI values
Acknowledgement
This project was funded by the U.S. Forest Service Forest
Health Monitoring Program (# SO-EM-F-10-01) and NASA
through a subcontract- WKU5162201000110070670 from
Western Kentucky University.
Figure 4. Classification and Regression Tree for the other hardwoods species group.
Contact Info: Michael K. Crosby (mkc34@msstate.edu) (662) 325-1527
Download